Backbone dynamics of Galectin-3C complexes : Bootstrap aggregation applied to spectral density function model selection
(2022) KEMR30 20221Department of Chemistry
- Abstract
- The spectral density function can be used to describe the stochastic motions of proteins. It has long been known that discrete values of the spectral density can be mapped by obtaining three Nuclear Magnetic Resonance (NMR) relaxation parameters: R1, R2 and NOE. In this work relaxation parameters are obtained at four different magnetic field strengths for two protein:ligand complexes of galectin-3C - a protein of great interest in cancer treatment research. Model-free model selection is performed to fit these parameters to various forms of the spectral density function. In addition to the conventional procedure, bootstrap aggregation (bagging) is implemented on model selection and parameter estimation which had distinct effects. The... (More)
- The spectral density function can be used to describe the stochastic motions of proteins. It has long been known that discrete values of the spectral density can be mapped by obtaining three Nuclear Magnetic Resonance (NMR) relaxation parameters: R1, R2 and NOE. In this work relaxation parameters are obtained at four different magnetic field strengths for two protein:ligand complexes of galectin-3C - a protein of great interest in cancer treatment research. Model-free model selection is performed to fit these parameters to various forms of the spectral density function. In addition to the conventional procedure, bootstrap aggregation (bagging) is implemented on model selection and parameter estimation which had distinct effects. The generalised order parameter is determined and it is shown that bagging has the potential to smooth out parameter values and reduce large differences. Furthermore, a new approach to adjust for excess contributions from chemical exchange without additional experimental procedures is investigated. (Less)
- Popular Abstract
- The world of proteins is unknown to most people enjoying their everyday lives. These molecules, too small to be seen with the bare eye, control our bodies; without them humans would not exist. Scientists have long studied proteins to, among other things, explain how we function, to understand deceases, to develop new drugs - and to create better washing powder. They are constantly seeking new ways of observing, manipulating and understanding proteins and some methods might be more unexpected than others. You know the small magnets you have sticking to your refrigerator; yes the ones you bought on your holiday to Mallorca last year? They can be used to study proteins. Well, not they precisely, but very, very strong magnets can. It might... (More)
- The world of proteins is unknown to most people enjoying their everyday lives. These molecules, too small to be seen with the bare eye, control our bodies; without them humans would not exist. Scientists have long studied proteins to, among other things, explain how we function, to understand deceases, to develop new drugs - and to create better washing powder. They are constantly seeking new ways of observing, manipulating and understanding proteins and some methods might be more unexpected than others. You know the small magnets you have sticking to your refrigerator; yes the ones you bought on your holiday to Mallorca last year? They can be used to study proteins. Well, not they precisely, but very, very strong magnets can. It might sound unbelievable, but it is actually the same technique as is used in Magnetic Resonance Imaging at hospitals - or MRI for short - which is likely a more familiar concept.
The function of proteins are closely related to their structure, hence many studies focus on finding a static shape. However, proteins are dynamic, which allows them to perform many of their vital tasks. In this work Nuclear Magnetic Resonance (NMR), which utilises incredibly strong electromagnets, is applied to investigate the dynamics of a protein called galectin-3C. More precisely, the movements of its backbone - the spine defining the overall structure of the protein - is examined. Galectin-3C is of interest as its functions are central in cancer cells and inflammation and new knowledge thus has the potential to break new grounds in the search towards treatments for these deceases.
The result of the NMR experiments is a set of three so called relaxation parameters. All experiments are performed on four different magnets (strong, very strong, super strong and strongest), thus 12 parameters are found in total. These parameters are then matched to various forms of a mathematical formula called the spectral density function, which can be used to describe the movements of a protein. We don't know which of the forms - which we call models - that is the best one, so we match the data to all of them and then determine which one to pick using statistical methods. Proteins can be divided into building blocks termed amino acid residues and conventionally one optimal model is found for each residue. Here, the challenge is approached a bit differently. Instead of selecting only a single model, all models are included in the final results, which are weighted after how well each model matches the data. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9093060
- author
- Nilsson, Daniel LU
- supervisor
-
- Mikael Akke LU
- organization
- course
- KEMR30 20221
- year
- 2022
- type
- H2 - Master's Degree (Two Years)
- subject
- keywords
- molecular biophysics, galectin, bootstrap aggregation, bagging, nuclear magnetic resonance, NMR
- language
- English
- id
- 9093060
- date added to LUP
- 2022-08-01 09:46:43
- date last changed
- 2022-08-01 09:46:43
@misc{9093060,
abstract = {{The spectral density function can be used to describe the stochastic motions of proteins. It has long been known that discrete values of the spectral density can be mapped by obtaining three Nuclear Magnetic Resonance (NMR) relaxation parameters: R1, R2 and NOE. In this work relaxation parameters are obtained at four different magnetic field strengths for two protein:ligand complexes of galectin-3C - a protein of great interest in cancer treatment research. Model-free model selection is performed to fit these parameters to various forms of the spectral density function. In addition to the conventional procedure, bootstrap aggregation (bagging) is implemented on model selection and parameter estimation which had distinct effects. The generalised order parameter is determined and it is shown that bagging has the potential to smooth out parameter values and reduce large differences. Furthermore, a new approach to adjust for excess contributions from chemical exchange without additional experimental procedures is investigated.}},
author = {{Nilsson, Daniel}},
language = {{eng}},
note = {{Student Paper}},
title = {{Backbone dynamics of Galectin-3C complexes : Bootstrap aggregation applied to spectral density function model selection}},
year = {{2022}},
}